Skip to content

Commit

Permalink
Add files via upload
Browse files Browse the repository at this point in the history
  • Loading branch information
inariksit authored Jun 16, 2020
1 parent 85b44e1 commit faec47e
Show file tree
Hide file tree
Showing 8 changed files with 127 additions and 0 deletions.
43 changes: 43 additions & 0 deletions cnl.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,43 @@
---
date: "2020-06-15"
---

# Controlled Natural Language

Based on real natural languages, but restricted in some way. (Paraphrased from [SIGCNL definition](https://www.sigcnl.org/index.html).)

## Types of restrictions

### Vocabulary
Example: only words related to seafaring or instruction manuals.

### Grammar
Example: don't allow [PP attachment](https://www.gabormelli.com/RKB/Prepositional_Phrase_Attachment_Task) ambiguity:

> I saw the man with a telescope
Variations [man with a telescope] and [saw with a telescope] need to be marked explicitly:

1. I saw the (man with a telescope)
1. I (saw with a telescope) the man

### Semantic

Example: TODO

## Purposes

### Improve communication among humans

Example: [Maritime English](https://en.wikipedia.org/wiki/Standard_Marine_Communication_Phrases)

People who work on ships need to learn a specific subset of English. When your ship hits an iceberg, you say:

> “I am fast in ice. I require assistance”.
instead of "the ship has crashed in ice, can somebody help me?"

This allows people to work on ships without knowing the full English language.
##
-
Some CNLs are designed to improve communication among humans, especially for non-native speakers of the respective natural language. In other cases, the restrictions on the language are supposed to make it easier for computers to analyze such texts in order to improve computer-aided, semi-automatic, or automatic translations into other languages. A third group of CNL has the goal to enable reliable automated reasoning and formal knowledge representation from seemingly natural texts.
7 changes: 7 additions & 0 deletions computable_contracts.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
---
date: "2020-06-15"
---

# Computable contracts

- <daml_smart_contracts>
11 changes: 11 additions & 0 deletions daml_smart_contracts.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@
---
date: "2020-06-15"
---

# DAML smart contracts

A different interpretation of contracts as code, where the contracts are
automatically generated and signed by computers and sent between servers.

For example, one company places an order of T-shirts for a conference. Now it
is important for both companies that they can prove that the order is placed and confirmed
4 changes: 4 additions & 0 deletions index.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,4 @@
# Overview

- <legal_drafting_is_broken>
- <computable_contracts>
23 changes: 23 additions & 0 deletions legal_drafting_is_broken.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
---
date: "2020-06-15"
---

# Legal drafting is broken

### Legal drafting uses suboptimal workers

Humans are imprecise:
* Incomplete information
* Semantically vague
* Syntactically ambiguous (e.g. <scope_ambiguity>)

### Legal drafting uses a suboptimal format
Plain text is lacking:
* Can't be checked for inconsistencies (except by humans who aren't good at it)
* Can't keep track of the overall state. If one thing changes, the rest needs to be updated manually.

Proposed solutions to fix legal drafting:

- <normalized_legal_drafting>
- <spreadsheets_for_law>
- <computable_contracts>
9 changes: 9 additions & 0 deletions normalized_legal_drafting.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,9 @@
---
date: "2020-06-15"
---

# Normalized legal drafting

1978 paper by Layman E. Allen _Normalized legal drafting and Query method_

Precursor to rules as code, computational contracts etc. "Normalized" legal document à la Allen ≈ <cnl>.
23 changes: 23 additions & 0 deletions scope_ambiguity.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
---
date: "2020-06-15"
---

# Scope ambiguity

Example from Allen & Engholm (1978) _Normalized Legal Drafting and the Query Method_

> Persons who are doctors and lawyers qualify.
What is the scope of _and_?

1. _[Persons who are doctors] and [persons who are lawyers] qualify_
1. Persons who are [doctors and lawyers] qualify


We will 100% need to support all kinds of scope ambiguity. Natural language
contracts are full of it. People like reading it and writing it. To spell out
"persons who are doctors qualify and persons who are lawyers qualify"
sounds like the worst kind of robotic <cnl>.

A full explicit sentence is useful for a human-directed ambiguity check
("did you mean … or …"), but not as the primary compilation target.
7 changes: 7 additions & 0 deletions spreadsheets_for_law.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
---
date: "2020-06-15"
---

# Spreadsheets for law

I don't know yet what this is, but Meng thinks it's important.

0 comments on commit faec47e

Please sign in to comment.